RRA failed
RRA failed
Dear all, I am now using RRA on arm26 model. As my simulation need, I have change the shoulder joint to a 6 dof customjoint just as the body pelvis in model gait2354.osim therefore I am able to use the RRA tool. I have referenced gaiat2354 to complete all the setupfiles including RRA. However, I once successfully compute the result but I forgot to save the file so when I tried to do it again, no matter how I try, the RRA turn out to have a weird result. During the RRA process, the whole arm just going downward and the joint just misplaces instead of going through the desired trajectory. What's more, I included a external force file with three force components and three torque which is sufficient fot the 6 dof customjoint to do RRA, however I can only process RRA with one of the dof locked. I have attached my file below if anyone could check is there anything wrong.
- Attachments
-
- arm_26.zip
- (33.32 KiB) Downloaded 3 times
-
- PIC1.png (397.37 KiB) Viewed 456 times
Tags:
-
Thomas Uchida
- Posts: 1800
- Joined: Wed May 16, 2012 11:40 am
Re: RRA failed
There's a red circle in the bottom-right corner of the application window. That means an error has occurred. You can get more information about the error by clicking the circle.
Re: RRA failed
Thanks for replying, the red circle appears in my others simulation so it's not related to the RRA simulations. I tried many times after I posted this post and I still don't know where's the problem is. The most frustrated thing is that I did it last week but fail this week while I thought I was doing everything the same.
-
Ton van den Bogert
- Posts: 167
- Joined: Thu Apr 27, 2006 11:37 am
Re: RRA failed
Gary,
We discussed this already in the other thread.
It's important to understand what 'residuals' are in the dynamic analysis. These are external forces and moments that are known to be zero.
In the arm26 model, even if you allow 6 dof in the shoulder joint, external forces and moments at the shoulder are needed to prevent the model from falling down under the influence of gravity. This makes sense, because you know there are physical structures between upper arm and ground (thorax), there is a shoulder joint, with ligaments and muscles. You can's assume zero force in those structures. These forces are there to prevent your arm from falling down.
In a gait model, external forces are provided by foot contact, and no external forces and moments are needed on the pelvis. And we know that those pelvis loads are zero, because there are no physical structures between pelvis and ground.
The arm model simply can't track your kinematic data without external forces and moments at the shoulder. Either the arm will fall down, or RRA will try to make the mass of the arm zero to prevent it from falling. Either way, that is a failure.
RRA should not be used on a system where external forces are needed. It will fail.
The good news is, RRA is not needed. I would go back to the original model and put a joint between shoulder and ground. The only reason for using RRA prior to CMC is to create dynamic consistency between model and motion, so that muscles can actually produce the motion. If you have nonzero external forces/moments acting on the pelvis, the model can, for instance, hang in the air and no amount of muscle force can do that. This why RRA is important in gait models.
In the arm model, there is no dynamic inconsistency so RRA is not needed. Any movement that you measure is dynamically feasible and can be produced by muscles (unless the muscles are too weak, which is a different problem).
Ton
We discussed this already in the other thread.
It's important to understand what 'residuals' are in the dynamic analysis. These are external forces and moments that are known to be zero.
In the arm26 model, even if you allow 6 dof in the shoulder joint, external forces and moments at the shoulder are needed to prevent the model from falling down under the influence of gravity. This makes sense, because you know there are physical structures between upper arm and ground (thorax), there is a shoulder joint, with ligaments and muscles. You can's assume zero force in those structures. These forces are there to prevent your arm from falling down.
In a gait model, external forces are provided by foot contact, and no external forces and moments are needed on the pelvis. And we know that those pelvis loads are zero, because there are no physical structures between pelvis and ground.
The arm model simply can't track your kinematic data without external forces and moments at the shoulder. Either the arm will fall down, or RRA will try to make the mass of the arm zero to prevent it from falling. Either way, that is a failure.
RRA should not be used on a system where external forces are needed. It will fail.
The good news is, RRA is not needed. I would go back to the original model and put a joint between shoulder and ground. The only reason for using RRA prior to CMC is to create dynamic consistency between model and motion, so that muscles can actually produce the motion. If you have nonzero external forces/moments acting on the pelvis, the model can, for instance, hang in the air and no amount of muscle force can do that. This why RRA is important in gait models.
In the arm model, there is no dynamic inconsistency so RRA is not needed. Any movement that you measure is dynamically feasible and can be produced by muscles (unless the muscles are too weak, which is a different problem).
Ton